Things are about to get worse for generative AI • Hacker News Board

Re: Things are about to get worse for generative AI

#91

Post by golol » Sat, Dec 30, 2023, 12:27 PM UTC

Baldbvrhunter wrote: post #2

I imagine the argument might be like this: I hire a session musician to play on my new single, paying him $100. I record the whole session. I ask him to play the opening to "Stairway to Heaven" and he does so. "Well, I can't use that as a sample without paying" "Ok play something like Jimmy Page" "Hmm, still sounds like Stairway to Heaven" "Ok, try and sound less like Stairway to Heaven but in that style" "Great, I'l…

If you release a media with copyrighted content it is IMO first and foremost your problem. Now if you have some contract with the guitarist that specifies that he produced a sample he has the rights to and sold it to you, but he clearly wasn't truthful, you can maybe pass the liability to him. This is not, however, how people will hse generative models If you use Dall-E you are not paying OpenAI to buy the rights to a piece Dall-E has produced. I see it more akin to hiring a musician to play for you for an hour, or a painter to paint for you. You are paying OpenAI to paint you something, but you I think OpenAI would never enter a contract which states that they are selling you the rights to a work.

Re: Things are about to get worse for generative AI

#92

Post by sjfjsjdjwvwvc » Sat, Dec 30, 2023, 12:28 PM UTC

rco8786 wrote: post #78

PartiallyTyped wrote: post #69

Earlier quoted context omitted.

creation should happen for its own sake. You don't see GMs stopping chess because bots are that much better.

creation competition. I agree with your premise but the chess analogy falls flat. We might, legitimately, see an enormous dropoff in people creating original works of literary, musical, and visual art (without AI).

This is an centuries old argument. Most people don’t create to make money, they create because they __have__ to.

If those motivated purely by money stop creating little of value will be lost

Re: Things are about to get worse for generative AI

#93

Post by golol » Sat, Dec 30, 2023, 12:30 PM UTC

geraldwhen wrote: post #75

docdeek wrote: post #60

How is this different to Googling “robot cop” or “video game plumber” and being served copyrighted material? Is it because Google will link to the image source? Or does the infringement begin when I use the image for gain, or claim it as my own? Perhaps it is because Google was allowed to crawl the page with the original image, so presenting them with a link is fine?

Looking at a copyrighted image posted by an author is not infringement. Printing that image onto a shirt and selling it is infringement. That’s what OpenAI is doing.

But OpenAI is not selling the rights to any images, or are they? When I pay for Dall-E, does the contract give me any rights for a work? If not then there is no issue.

Re: Things are about to get worse for generative AI

#94

Post by AlexAndScripts » Sat, Dec 30, 2023, 12:30 PM UTC

sanderjd wrote: post #67

meindnoch wrote: post #57

Earlier quoted context omitted.

Developing AGI is a matter of national security. "Content" isn't.

I'm curious: Have you seen indications that major militaries and politicians believe AGI , rather than special purpose ML for military purposes, is important for national security? I'm really not sure whether this is true, or whether military and political leaders think it's true.

They will once people start talking about an AI gap with China.

Re: Things are about to get worse for generative AI

#95

Post by RandomLensman » Sat, Dec 30, 2023, 12:31 PM UTC

taberiand wrote: post #85

gchamonlive wrote: post #62

Earlier quoted context omitted.

The thing is that those are really trivial or extreme examples. What we should take from this: 1. Generative AI systems are fully capable of producing materials that infringe on copyright. 2. They do not inform users when they do so. So potentially any output could be infringing copyright source material, even from some obscure but still protected corner of the web, and anyone using that output could be exposed to la…

Why isn't the solution a strong disclaimer when the user generates images along the lines of "beware that the images produced may not comply with copyright laws in your country" etc etc? Any artist can privately draw a picture of Mario, what's so different about having an LLM generate that image?

There is a huge amount of copyrighted material in the world and these tools let you create a lot of content in a short time, that would imply an increased risk vs slower, more manual approaches.

What I am also wondering is, if for some things copyrighted material could be somehow dominant in the output (beyond prompting more or less for it).

Re: Things are about to get worse for generative AI

#96

Post by golol » Sat, Dec 30, 2023, 12:31 PM UTC

nojs wrote: post #68

In practice, what happens next when websites all start to block openai by default (or change their TOS to disallow OpenAI’s crawlers)? It seems like there’s little incentive not to do this, because unlike Google OpenAI isn’t bringing any traffic or eyeballs. It may end up being a default setting in Wordpress for example. But OpenAI presumably can’t afford to pay every single long tail source of content on the whole i…

It's not like you can hide the web from OpenAI. They could just use a secret crawler. Or buy the data from a third party company.

Re: Things are about to get worse for generative AI

#97

Post by gchamonlive » Sat, Dec 30, 2023, 12:32 PM UTC

keiferski wrote: post #79

gchamonlive wrote: post #62

Earlier quoted context omitted.

The thing is that those are really trivial or extreme examples. What we should take from this: 1. Generative AI systems are fully capable of producing materials that infringe on copyright. 2. They do not inform users when they do so. So potentially any output could be infringing copyright source material, even from some obscure but still protected corner of the web, and anyone using that output could be exposed to la…

But how is that any different from creating an image from scratch? If I make a logo and use it for my business, but it turns out to be very similar to one already being used by another company, it’s the same situation. I think the main concern here is with the top 1,000 or so brands/copyrights which seem fairly straightforward to deal with using the method I described.

It's plagiarism (https://www.youtube.com/watch?v=yDp3cB5fHXQ).

It's not the same situation. You can't possibly expect someone to be exposed to the entirety of the internet like ChatGPT is. It is a matter of scale. If you still think they are the same thing, the industrial revolution was about scale and had transformative impacts in the society.

Re: Things are about to get worse for generative AI

#98

Post by disgruntledphd2 » Sat, Dec 30, 2023, 12:32 PM UTC

davidy123 wrote: post #80

The solution could be great. I really don't like the way culture always goes to the same tropes, calling any potential innovation "out of Star Trek" (with attendant distorted expectations), right down to expecting an interface based on literal hand-waving in Minority Report. If copyright held works ("USS Enterprise") could be removed, yet the actual essential concepts (space ship, naming things) retained, it would be…

Science journals are mostly under copyright of a few big publishers who are extremely hostile to any kind of ML being performed on the content.

Re: Things are about to get worse for generative AI

#99

Post by rolisz » Sat, Dec 30, 2023, 12:33 PM UTC

Simple fix (at least for ChatGPT): ask it to avoid drawings with similarities to copyrighted characters.

Re: Things are about to get worse for generative AI

#100

Post by keiferski » Sat, Dec 30, 2023, 12:34 PM UTC

mrweasel wrote: post #72

keiferski wrote: post #39

These don't seem all that difficult to fix to me. Most of the examples are not really generic, but are shorthand descriptions of well-known entities. "Video game plumber" is practically synonymous with "Mario" and anyone that has the slightest familiarity with the character knows this. Likewise, how difficult is it to just use descriptive tools to describe Mario-like images [1] and then remove these results from anyo…

It's going to be hard to remove every single "shorthand descriptions of well-known entities" or other prompts that can be used to generate copyrighted or trademarked content. Sure, if you're not deliberately trying to generate infringing content, you can probably remove or discard those results, the trouble is the people who will try to trick the AI to generate this content, blocking those people is going to be impos…

But does this actually matter if the people are only generating images for their own use? Does Photoshop prevent people from making drawings that look like Mickey Mouse? Of course not.

I think it will be easy to prevent the obvious copyrighted stuff via the method I mentioned. People going around those restrictions are subject to the same rules as someone drawing the copyrighted image from scratch.